wild deer, so that the soldiers and sailors had plenty of food.” My own impression on the spot was that Ammianus Marcellinus’s Dura must be Irzî. The tower tombs were certainly erected before the middle of the fourth century, therefore they were in existence when Julian passed; moreover, they were far more numerous and conspicuous than they are at present, since almost all of them have now fallen into ruin. It is difficult to see how Irzî could have failed to attract the attention of Ammianus Marcellinus, and Dura is the one place mentioned by him between Zeitha and ’Ânah. But the Dura of Isidorus, the nineteenth station, has to be placed at Abu’l Ḥassan, not at Irzî, since his twentieth station, Merrhan, necessarily falls at Irzî, and I can only conjecture that, as in Julian’s time both places were ruined and deserted, Ammianus Marcellinus made a confusion between them, or was wrongly informed, and transferred the name of Dura (Abu’l Ḥassan) to Merrhan (Irzî). For the twenty-first station, Giddan, I can offer no suggestion. Jabarîyeh will scarcely fit, as it is but 13 hrs. 15 min. from ’Ânah, and Giddan was 17 sch. from Anatho, but it must be admitted that all the distances between the stations from Merrhan to ’Ânah seem to be too long according to my caravan time. The twenty-second station, Belesibiblada, was placed by Chesney at Ḳal’at Bulâḳ, and I saw no better site for it, though I took only 9 hrs. and 25 min. to reach it from Irzî, and the distance given by Isidorus is 12 sch. Ritter would place at Ḳal’at Bulâḳ Ptolemy’s Bonakhe. I do not see any way of identifying with certainty the island station, the twenty-third, which was 4 sch. from ’Ânah. There are many islands in the stream above ’Ânah. One of them, Ḳarâbileh, is reported to have ruins upon it; it was about four hours’ journey from ancient ’Ânah, and may therefore be identical with the twenty-third station, which is placed at a distance of 4 sch. from Anatho. Anatho, the twenty-fourth station, Isidorus expressly states to be on an island; it was therefore the successor to the Assyrian fortress which I believe to have existed on the island of Lubbâd. Xenophon does not mention it; nor does Ptolemy, unless his Bethanua may be taken for ’Ânah as Ritter believed (Vol. XI. p. 716). Rawâ may possibly be the Phathusa of Zosimos, but I would rather place Phathusa on the left bank, opposite and below the island of Lubbâd, where there are many mounds and ruins. I did not follow the river below ’Ânah very closely, but the ruins I saw near Ḥadîthah help to justify the presumption that Olabus was situated there. Chesney wished to identify Izannesopolis with the ruins of a castle between Baghdâdî and Hît. I did not go to the spot, and my caravan time between Ḥadîthah and Hît is therefore rather misleading, for if I had followed the river so as to visit the ḳaṣr, the journey would have taken more than the seventeen and a half hours which I have recorded. Isidorus’s 16 sch.